A Preference-Aware Service Recommendation Method on Map-Reduce

Service recommender systems have shown to be valuable tools to provide appropriate recommendations to the users. However, in most of existing service recommender systems, the ratings and rankings of services presented to different users are the same, which didn't consider users' preferences and therefore could not meet users' personalized requirements. Moreover, the number of customers, alternative services and other online information grows rapidly. Thus, the improvement of scalability and efficiency of recommender systems is also necessary and urgent. In view of these challenges, a preference-aware service recommendation method on Map-Reduce, named PASR, is proposed in this paper. It aims at presenting a personalized ranking list and recommending the most appropriate services to the users from big data environment. In this method, keywords are used to indicate users' preferences, and a user-based Collaborative Filtering algorithm is adopted to generate appropriate recommendations. To improve the scalability and efficiency of PASR, we implement it on a distributed computing platform, Hadoop, which uses Map-Reduce as its computing framework. Finally, experimental results show that our approach performs well both in accuracy and scalability.